EP1967589A2

EP1967589A2 - An oleosin 5' regulatory region for the modification of plant seed lipid composition

Info

Publication number: EP1967589A2
Application number: EP08155814A
Authority: EP
Inventors: Zhongsen Li; Terry Thomas
Original assignee: Bayer CropScience SA
Current assignee: Bayer SAS
Priority date: 1997-04-09
Filing date: 1998-04-09
Publication date: 2008-09-10
Anticipated expiration: 2018-04-09
Also published as: EP1967589B1; CN1259168A; CA2285687A1; DE69839667D1; ES2398603T3; ATE399870T1; ES2307320T3; AU7107198A; WO1998045461A1; BR9807969A; AR012399A1; US5977436A; PT973920E; DK0973920T3; JP2001519668A; EP0973920B1; ZA983047B; CA2285687C; EP1967589A3; AU739442B2

Abstract

The present invention is directed to 5' regulatory regions of an Arabidopsis oleosin gene.
The 5' regulatory regions, when operably linked to either the coding sequence of a gene or a sequence complementary to a native plant gene, direct expression of the coding sequence or complementary sequence in a plant seed. The regulatory regions are useful in expression cassettes and expression vectors for the transformation of plants. Also provided are methods of modulating the levels of a heterologous gene such as a fatty acid synthesis or lipid metabolism gene by transforming a plant with the subject expression cassettes and expression vectors.

Description

BACKGROUND OF THE INVENTION

Seed oil content has traditionally been modified by plant breeding. The use of recombinant DNA technology to alter seed oil composition can accelerate this process and in some cases alter seed oils in a way that cannot be accomplished by breeding alone, The oil composition of Brassica has been significantly altered by modifying the expression of a number of lipid metabolism genes. Such manipulations of seed oil composition have focused on altering the proportion of endogenous component fatty acids. For example, antisense repression of the Δ12-desaturase gene in transgenic rapeseed has resulted in an increase in oleic acid of up to 83%. Topfer et al. 1995 Science 268:681-686.
There have been some successful attempts at modifying the composition of seed oil in transgenic plants by introducing new genes that allow the production of a fatty acid that the host plants were not previously capable of synthesizing. Van de Loo, et al. (1995 Proc. Natl. Acad. Sci USA 92:6743-6747) have been able to introduce a Al2-hydroxyla.se gene into transgenic tobacco, resulting in the introduction of a novel fatty acid, ricinoleic acid, into its seed oil. The reported accumulation was modest from plants carrying constructs in which transcription of the hydroxylase gene was under the control of the cauliflower mosaic virus (CaMV) 35S promoter, Similarly, tobacco plants have been engineered to produce low levels of petroselinic acid by expression of an acyl-ACP desaturase from coriander (Cahoon et al. 1992 Proc. Natl. Acad, Sci USA 89: 11184-11188).
The long chain fatty acids (C18 and larger), have significant economic value both as nutritionally and medically important foods and as industrial commodities (Ohlrogge.J.B. 1994 Plant Physiol. 104:811-826). Linoleic (18:2 Δ9,12) and α-linolenic acid (18:3 Δ9,12,15) are essential fatty acids found in many seed oils. The levels of these fatty-acids have been manipulated in oil seed crops through breeding and biotechnology (Ohlrogge, et al 1991 Biochim. Biophis. Acta 1082:1-26; Topfer et al. 1995 Science 268:681-686). Additionally, the production of novel fatty acids in seed oils can be of considerable use in both human health and industrial applications.
Consumption of plant oils rich in γ-linolenic acid (GLA) (18:3 Δ6,9,12) is thought to alleviate hypercholesterolemia and other related clinical disorders which correlate with susceptibility to coronary heart disease (Brenner R.R. 1976 Adv. Exp, Med. Biol 83:85-101) . The therapeutic benefits of dietary GLA may result from its role as a precursor to prostaglandin synthesis (Weete, J.D. 1980 in Lipid Biochemistry of Fungi and Other Organisms, eds. Plenum Press, New York, pp. 59-62). Linoleic acid(18:2) (LA) is transformed into gamma linolenic acid (18.3) (GLA) by the enzyme Δ6-desaturase.
Few seed oils contain GLA despite high contents of the precursor linoleic acid. This is due to the absence of Δ6-desaturase activity in most plants. For example, only borage (Borage officinalis), evening primrose (Oenothera biennis), and currants (Ribes nigrum) produce appreciable amounts of linolenic acid. Of these three species, only Genothera and Borage are cultivated as a commercial source for GLA. It would be beneficial if agronomic seed oils could be engineered to produce GLA in significant quantities by introducing a heterologous Δ6-desaturase gene. It would also be beneficial if other expression products associated with fatty acid synthesis and lipid metabolism could be produced in plants at high enough levels so that commercial production of a particular expression product becomes feasible.
As disclosed in U.S. Patent No. 5,552,306 , a cyanobacterial Δ6-desaturase gene has been recently isolated. Expression of this cyanobacterial gene in transgenic tobacco resulted in significant but low level GLA accumulation. (Reddy et al, 1996 Nature Biotech. 14:639-642). Applicant's copending U.S. Application Serial No. 08,366,779 , discloses a Δ6-desaturase gene isolated from the plant Borago officinallis and its expression in tobacco under the control of the CaMV 35S promoter. Such expression resulted in significant but low level GLA and octadecatetraenoic acid (ODTA. or OTA) accumulation in seeds, Thus, a need exists for a promoter which functions in plants and which consistently directs high level expression of lipid metabolism genes in transgenic plant seeds.
Oleosins are abundant seed proteins associated with the phospholipid monolayer membrane of oil bodies. The first oleosin gene, L3, was cloned from maize by selecting clones whose in vitro translated products were recognized by an anti-L3 antibody (Vance et a.l. 1987 J. Biol. Chem. 262:11275-11279 ).Subsequently, different isoforms of oleosin genes from such different species as Brassica, soybean, carrot, pine, and Arabidopsis have been cloned (Huang, A.H.C., 1992, Ann. Reviews Plant Phys. and Plant Mol. Biol. 43:177-200; Kirik et al., 1996 Plant Mol. Biol. 31:413-417; Van Rooijen et al., 1992 Plant Mol. Biol. 18: 1177-1179 ; Zou et al. , Plant Mol. Biol . 31:429-433. Oleosin protein sequences predicted from these genes are highly conserved, especially for the central hydrophobic domain. All of these oleosins have the characteristic feature of three distinctive domains. An amphipathic domain of 40-60 amino acids is present at the N-termimus ; a totally hydrophobic domain of 68-74 amino acids is located at the center; and an amphipathic α-helical domain of 33-40 amino acids is situated at the C-terminus (Huang A.H.C. 1992).
The present invention provides 5' regulatory sequences from an oleosin gene which direct high level expression of lipid metabolism genes in transgenic plants. In accordance with the present invention, chimeric constructs comprising an oleosin 5' regulatory region operably linked to coding sequence for a lipid metabolism gene such as a Δ6-desaturase gene are provided. Transgenic plants comprising the subject chimeric constructs produce levels of GLA approaching the level found in those few plant species which naturally produce GLA such as evening primrose (Oenothera biennis).

SUMMARY OF THE INVENTION

The present invention is directed to 5' regulatory regions of an Arabidopsis oleosin gene. The 5' regulatory regions, when operably linked to either the coding sequence of a heterologous gene or sequence complementary to a native plant gene, direct expression of the heterologous gene or complementary sequence in a plant seed.
The present invention, thus provides expression cassettes and expression vectors comprising an oleosin 5' regulatory region operably linked to a heterologous gene or a sequence complementary to a native plant gene.
Plant transformation vectors comprising the expression cassettes and expression vectors are also provided as are plant cells transformed by these vectors, and plants and their progeny containing the vectors,
In one embodiment of the invention, the heterologous gene or complementary gene sequence is a fatty acid synthesis gene or a lipid metabolism gene.
In another aspect of the present invention, a method is provided for producing a plant with increased levels of a product of a fatty acid synthesis or lipid metabolism gene.
In particular, there is provided a method for producing a plant with increased levels of a fatty acid synthesis or lipid metabolism gene by transforming a plant with the subject expression cassettes and expression vectors which comprise an oleosin 5' regulatory region and a coding sequence for a fatty acid synthesis or lipid metabolism gene.
In another aspect of the present invention, there is provided a method for cosuppressing a native fatty acid synthesis or lipid metabolism gene by transforming a plant with the subject expression cassettes and expression vectors which comprise an oleosin 5' regulatory region and a coding sequence for a fatty acid synthesis or lipid metabolism gene.
A further aspect of this invention provides a method of decreasing production of a native plant gene such as a fatty acid synthesis gene or a lipid metabolism gene by transforming a plant with an expression vector comprising a oleosin 5' regulatory region operably linked to a nucleic acid sequence complementary to a native plant gene.
Also provided are methods of modulating the levels of a heterologous gene such as a fatty acid synthesis or lipid metabolism gene by transforming a plant with the subject expression cassettes and expression vectors.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 depicts the nucleotide and corresponding amino acid sequence of the borage Δ6-desaturase gene (SEQ ID NO:1). The cytochrome b5 heme-binding motif is boxed and the putative, metal binding, histidine rich motifs (HRMs) are underlined. The motifs recognized by the primers (PCR analysis) are underlined dotted lines, i.e. tgg aaa tgg aac cat aa; and gag cat cat ttg ttt cc.
Fig. 2 is a dendrogram showing similarity of the borage Δ6-desaturas to other membrane-bound desaturases. The amino acid sequence of the borage Δ6-desaturase was compared to other known desaturases using Gene Works (IntelliGenetics). Numerical values correlate to relative phylogenetic distances between subgroups compared.
Fig. 3A provides a gas liquid chromatography profile of the fatty acid methyl esters (FAMES) derived from leaf tissue of a wild type tobacco 'Xanthi'.
Fig. 3B provides a gas liquid chromatography profile of the FAMES derived from leaf tissue of a tobacco plant transformed with the borage Δ6-desaturase cDNA under transcriptional control of the CaMV 35S promoter (pAN2). Peaks corresponding to methyl linoleate (18:2), methyl γ-linolenate (18:3γ), methyl α-linolenate (18:3α), and methyl octadecatetraenoate (18:4) are indicated.
Fig. 4 is the nucleotide sequence and corresponding amino acid sequence of the oleosin AtS21 cDNA (SEQ ID NO: 3)
Fig. 5 is an acidic-base map of the predicted AtS21 protein generated by DNA Strider 1.2,
Fig. 6 is a Kyte-Doolittle plot of the predicted AtS21 protein generated by DNA Strider 1.2.
Fig. 7 is a sequence alignment of oleosins isolated from Arabidopsis. Oleosin sequences published or deposited in EMBL, BCM, NCBI databases were aligned to each other using GeneWorks^® 2.3. Identical residues are boxed with rectangles. The seven sequences fall into three groups. The first group includes AtS21 (SEQ ID NO:5), X91918 (SEQ ID NO:6) and Z29859 (SEQ ID NO:7). The second group includes X62352 (SEQ ID NO:8) and Ato13 (SEQ ID NO:9) . The third group includes X91956 (SEQ ID NO:10) and L40954 (SEQ ID NO:11). Differences in amino acid residues within the same group are indicated by shadows, Ato2/Z54164 is identical to AtS21. Atol.3 sequence (Accession No. Z541654 in EMBL database) is actually not disclosed in the EMBL database, The Z54165 Accession number designates the same sequence as Z54164. which is Atol2.
Fig. 8A is a Northern analysis of the AtS21 gene. An RNA gel blot containing ten micrograms of total RNA extracted from Arabidopsis flowers (F), leaves (L), roots (R), developing seeds (Se), and developing silique coats (Si) was hybridized with a probe made from the full-length AtS21 cDNA.
Fig. 8B is a Southern analysis of the AtS21 gene. A DNA gel blot containing ten micrograms of genomic DNA digested with BamHI (B), EcoRI (E), HindIII (H), SacI (S), and XbaI (X) was hybridized with a probe made from the full length AtS21 cDNA.
Fig. 9 is the nucleotide sequence of the SacI fragment of AtS21 genomic DNA. (SEQ ID NO:12) , The promoter and intron sequences are in uppercase. The fragments corresponding to AtS21 cDNA sequence are in lower case- The first ATG codon and a putative TATA box are shadowed. The sequence complementary to 21P primer for PCR amplification is boxed. A putative abscisic acid response element (ABRE) and two 14 bp repeats are underlined.
Fig. 10 is a map of AtS21 promoter/GUS construct (PAN5).
Fig. 11A depicts AtS21/GUS gene expression in Arabidopsis bolt and leaves.
Fig. 11B depicts AtS21 GUS gene expression in Arabidopsis siliques.
Fig. 11C depicts AtS21 GUS gene expression in Arabidopsis developing seeds.
Figs. 11D through 11J depict AtS21 GUS gene expression in Arabidopsis developing embryos,
Fig. 11K depicts AtS21/GUS gene expression in Arabidopsis root and root hairs of a young seedling.
Fig. 11L depicts Ats21/GUS: gene expression in Arabidopsis cotyledons and the shoot apex of a five day seedling.
Figs. 11M and 11N depict AtS21/GUS gene expression in cotyledons and the shoot apex of 5-15 day seedlings.
Fig. 12A depicts AtS21/GUS gene expression in tobacco embryos and endosperm.
Fig. 12B depicts AtS21/GUS gene expression in germinating tobacco seeds.
Fig. 12C depicts AtS21/GUS gene expression in a day old tobacco seedling.
Fig. 12D depicts AtS21/GUS gene expression in 5-15 day old tobacco seedlings.
Fig. 13A is a Northern analysis showing A1S21 mRNA levels in developing wild-type Arabidopsis seedlings, Lane 1 was loaded with RNA from developing seeds, lane 2 was loaded with RNA from seeds imbibed for 24-48 hours, lane 3: 3 day seedlings; lane 4: 4 day seedlings; lane 5: 5 day seedlings; lane 6: 6 day seedlings; lane 7; 9 day seedlings; lane 8: 12 day seedlings. Probe was labeled AtS21 cDNA. Exposure was for one hour at -80°C.
Fig. 13B is the same blot as Fig, 13A only exposure was for 24 hours at -80°C.
Fig. 13C is the same blot depicted in Figs. 13A and 13B after stripping and hybridization, with an Arabidopsis tubulin gene probe. The small band in each of lanes 1 and 2 is the remnant of the previous At S21 probe. Exposure was for 48 hours at -80°C.
Fig. 14 is a graph comparing GUS activities expressed by the AtS21 and 35S promoters. GUS activities expressed by the AtS21 promoter in developing Arabidopsis seeds and leaf are plotted side by side with those expressed by the 35S promoter. The GUS activities expressed by the AtS21 promoter in tobacco dry seed and leaf are plotted on the right side of the figure. GUS activity in tobacco leaf is so low that no column appears. «G-H» denotes globular to heart stage; «H-T» denotes heart to torpedo stage; «T-C» denotes torpedo to cotyledon stage; «Early C» denotes early cotyledon; «Late C» denotes late cotyledon, The standard deviations are listed in Table 2.
Fig. 15A is an RNA gel blot analysis carried out on 5 µg samples of RNA isolated from borage leaf, root, and 12 dpp embryo tissue, using labeled borage Δ6-desaturase cDNA as a hybridization probe.
Fig. ISB depicts a graph corresponding to the Northern analysis results for the experiment shown in Fig. 15A.
Fig. 16A is a graph showing relative legumin RNA accumulation in developing borage embryos based on results of Northern blot.
Fig. 16B is a graph showing relative oleosin RNA accumulation in developing borage embryos based on results of Northern blot.
Fig. 16C is a graph showing relative Δ6-desaturase RNA accumulation in developing borage embryos based on results of Northern blot.
Fig. 17 is a PCR analysis showing the presence of the borage delta 6-desaturase gene in transformed plants of oilseed rape. Lanes 1, 3 and 4 were loaded with PCR reactions performed with DMA from plants transformed with the borage delta 6-desaturase gene linked to the oleosin 5' regulatory region; lane 2: DNA from plant transformed with the borage delta 6-desaturase gene linked to the albumin 5' regulatory region; lanes 5 and 6: DNA from non-transformed plants, lane 7: molecular weight marker (1 kb ladder, Gibco BRL); lane 8: PCR without added template DNA; lane 9: control with DNA from Agrobacterium tumefaciens EHA 105 containing the plasmid pAN3 (i.e. the borage delta6-desaturase gene linked to the oleosin 5 regulatory region).

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides isolated nucleic acids encoding 5' regulatory regions from an Arabidopsis oleosin gene. In accordance with the present invention, the subject 5' regulatory regions, when operably linked to either a coding sequence of a heterologous gene or a sequence complementary to a native plant, gene, direct expression of the coding sequence or complementary sequence in a plant seed. The oleosin 5' regulatory regions of the present invention are useful in the construction of an expression cassette which comprises in the 5' to 3' direction, a subject oleosin 5' regulatory region, a heterologous gene or sequence complementary to a native plant gene under control of the regulatory region and a 3' termination sequence. Such an expression cassette can be incorporated into a variety of autonomously replicating vectors in order to construct an expression vector.
It has been surprisingly found that plants transformed with the expression vectors of the present invention produce levels of GLA approaching the level found in those few plant species which naturally produce GL4 such as evening primrose (Oenothera biennis).
As used herein, the term «cassette» refers to a nucleotide sequence capable of expressing a particular gene if said gene is inserted so as to be operably linked to one or more regulatory regions present in the nucleotide sequence. Thus, for example, the expression cassette may comprise a heterologous coding sequence which is desired to be expressed in a plant seed. The expression cassettes and expression vectors of the present invention are therefore useful for directing seed-specific expression of any number of heterologous genes. The term «seed-specific expressions as used herein, refers to expression in various portions of a plant seed such as the endosperm and embryo.
An isolated nucleic acid encoding a 5' regulatory region from an oleosin gene can be provided as follows. Oleosin recombinant genomic clones are isolated by screening a plant genomic DNA library with a cDNA (or a portion thereof) representing oleosin mRNA. A number of different oleosin cDNAs have been isolated. The methods used to isolate such cDNAs as well as the nucleotide and corresponding amino acid sequences have been published in Kirik et al. 1986 Plant Mol. Biol. 31:413-417; Zou et al. Plant Mol. Biol. 31:429-433; Van Rooigen et al. 1992 Plant Mol. biol 18: 1177-1179.
Virtual subtraction screening of a tissue specific library using a random primed polymerase chain (RP-PCR) cDNA probe is another method of obtaining an oleosin cDNA useful for screening a plant genomic DNA library, Virtual subtraction screening refers to a method where a cDNA library is constructed from a target tissue and displayed at a low density so that individual cDNA clones can be easily separated. These cDNA clones are subtractively screened with driver quantities (i.e., concentrations of DNA to kinetically drive the hybridization reaction) of cDNA probes made from tissue or tissues other than the target tissue (i.e. driver tissue). The hybridized plaques represent genes that are expressed in both the target and the driver tissues; the unhybridized plaques represent genes that may be target tissue-specific or low abundant genes that can not be detected by the driver cDNA probe. The unhybridized cDNAs are selected as putative target tissue-specific genes and further analyzed by one-pass sequencing and Northern hybridization.
Random primed PCR (RP-PCR) involves synthesis of large quantities of cDNA probes from a trace amount of cDNA template. The method combines the amplification power of PCR with the representation of random priming to simultaneously amplify and label double-stranded cDNA in a single tube reaction.
Methods considered useful in obtaining oleosin genomic recombinant DNA are provided in Sambrook et al. 1989, in Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY, for example, or any of the myriad of laboratory manuals on recombinant DNA technology that are widely available. To determine nucleotide sequences, a multitude of techniques are available and known to the ordinarily skilled artisan. For example, restriction, fragments containing an oleosin regulatory region can be subcloned into the polylinker site of a sequencing vector such as pBluescript (Stratagene). These pBluescript subclones can then be sequenced by the double-stranded dideoxy method (Chen and Seeburg, 1985, DNA 4:165).
In a preferred embodiment, the oleosin regulatory region comprises nucleotides 1-1267 of Fig. 9 (SEQ ID NO: 12). Modifications to the oleosin regulatory region as set forth in SEQ ID NO:12 which maintain the characteristic property of directing seed-specific expression, are within the scope of the present invention. Such modifications include insertions, deletions and substitutions of one or more nucleotides.
The 5' regulatory region of the present invention can be derived from restriction endonuclease or exonuclease digestion of an oleosin genomic clone, Thus, for example, the known nucleotide or amino acid sequence of the coding region of an isolated oleosin gene (e.g. Fig. 7) is aligned to the nucleic acid or deduced amino acid sequence of an isolated oleosin genomic clone and 5' flanking sequence (i.e., sequence upstream from the translational start codon of the coding region) of the isolated oleosin genomic clone located.
The oleosin 5' regulatory region as set forth in SEQ ID NO:12 (nucleotides 1-1267 of Fig. 9) may be generated from a genomic clone having either or both excess 5' flanking sequence or coding sequence by exonuclease III-mediated deletion. This is accomplished by digesting appropriately prepared DNA with exonuclease III (exoIII) and removing aliquots at increasing intervals of time during the digestion. The resulting successively smaller fragments of DNA may be sequenced to determine the exact endpoint of the deletions. There are several commercially available systems which use exonuclease III (exoIII) to create such a deletion series, e.g. Promega Biotech, «Erase - A-Ease» system. Alternatively, PCR primers can be defined to allow direct amplification of the subject 5' regulatory regions.
Using the same methodologies, the ordinarily skilled artisan can generate one or more deletion fragments of nucleotides 1-1267 as set forth in SEQ ID NO:12. Any and all deletion fragments which comprise a contiguous portion of nucleotides set forth in SEQ ID NO:12 and which retain the capacity to direct seed-specific expression are contemplated by the present invention,
The identification of oleosin 5' regulatory sequences which direct seed-specific expression comprising nucleotides 1-1267 of SEQ ID NO:12 and modifications or deletion fragments thereof, can be accomplished by transcriptional fusions of specific sequences with the coding sequences of a heterologous gene, transfer of the chimeric gene into an appropriate host, and detection of the expression of the heterologous gene The assay used to detect expression depends upon the nature of the heterologous sequence. For example, reporter genes, exemplified by chloramphenicol acetyl transferase and β-glucuronidase (GUS), are commonly used to assess transcriptional and translational competence of chimeric constructions. Standard assays are available to sensitively detect the reporter enzyme in a transgenic organism. The β-glucuroxiidase (GUS) gene is useful as a reporter of promoter activity in transgenic plants because of the high stability of the enzyme in plant cells, the lack of intrinsic β-glucuroni-dase activity in higher plants and availability of a quantitative fluorimetric assay and a histochemical localization technique. Jefferson et al. (1987 EMBO J 6:3901) have established standard procedures for biochemical and histochemical detection of GUS activity in plant tissues. Biochemical assays are performed by mixing plant tissue lysates with 4-methylumbelliferyl-β-D-glucuronide, a fluorimetric substrate for GUS, incubating one hour at 37°C, and then measuring the fluorescence of the resulting 4-methyl-umbelliferone. Histochemical localization for GUS activity is determined by incubating plant tissue samples in 5-bromo-4-chloro-3-indolyl-glucuronide (X-Glue) for about 18 hours at 37°C and observing the staining pattern of X-Gluc. The construction of such chimeric genes allows definition of specific regulatory sequences and demonstrates that these sequences can direct expression of heterologous genes in a seed-specific manner.
Another aspect of the invention is directed to expression cassettes and expression vectors (also termed herein «chimeric genes») comprising a 5' regulatory region from an oleosin gene which directs seed specific expression operably linked to the coding sequence of a heterologous gene such that the regulatory element is capable of controlling expression of the product encoded by the heterologous gene, The heterologous gene can be any gene other than oleosin. If necessary, additional regulatory elements or parts of these elements sufficient to cause expression resulting in production of an effective amount of the polypeptide encoded by the heterologous gene are included in the chimeric constructs,
Accordingly, the present invention provides chimeric genes comprising sequences of the oleosin 5' regulatory region that confer seed-specific expression which are operably linked to a sequence encoding a heterologous gene such as a lipid metabolism enzyme. Examples of lipid metabolism genes useful for practicing the present invention include lipid desaturases such as Δ6-desaturases, Δ12-desaturases, Δ15-desaturases and other related desaturases such as stearoyl-ACP desaturases, acyl carrier proteins (ACPs), thioesterases, acetyl transacylases, acetyl-coA carboxylases, ketoacyl-synthases, malonyl transacylases, and elongases. Such lipid metabolism genes have been isolated and characterized from a number of different bacteria and plant species. Their nucleotide coding sequences as well as methods of isolating such coding sequences are disclosed in the published literature and are widely available to those of skill in the art.
In particular, the Δ6-desaturase genes disclosed in U.S. Patent No. 5,552,306 and applicants' copending U.S. Application serial No. 08/366,779 filed December 30, 1994 and incorporated herein by reference, are contemplated as lipid metabolism genes particularly useful in the practice of the present invention.
The chimeric genes of the present invention are constructed by ligating a 5' regulatory region of a oleosin genomic DNA to the coding sequence of a heterologous gene, The juxtaposition of these sequences can be accomplished in a variety of ways. In a preferred embodiment the order of the sequences, from 5' to 3', is an oleosin 5' regulatory region (including a promoter) a coding sequence, and a termination sequence which includes a polyadenylation site.
Standard techniques for construction of such chimeric genes are well known to those of ordinary skill in the art and can be found in references such as Sambrook et al.(1989), A variety of strategies are available for ligating fragments of DNA, the choice of which depends on the nature of the termini of the DNA fragments, One of ordinary skill in the art recognizes that in order for the heterologous gene to be expressed, the construction requires promoter elements and signals for efficient polyadenylation of the transcript. Accordingly, the oleosin 5' regulatory region that contains the consensus promoter sequence known as the TATA box can be ligated directly to a promoterless heterologous coding sequence.
The restriction or deletion fragments that contain the oleosin TATA box are ligated in a forward orientation to a promoterless heterologous gene such as the coding sequence of fi--glucuronidase {GUS}. The skilled artisan will recognize that the subject oleosin 5' regulatory regions can be provided by other means, for example chemical or enzymatic synthesis. The 3' end of a heterologous coding sequence is optionally ligated to a termination sequence comprising a polyadenylation site, exemplified by, but not limited to, the nopaline synthase polyadenylation site, or the octopine T-DNA gene 7 polyadenylation site. Alternatively, the polyadenylation site can be provided by the heterologous gene.
The present invention also provides methods of increasing levels of heterologous genes in plant seeds. In accordance with such methods, the subject expression cassettes and expression vectors are introduced into a plant in order to effect expression of a heterologous gene. For example, a method of producing a plant with increased levels of a product of a fatty acid synthesis or lipid metabolism gene is provided by transforming a plant cell with an expression vector comprising an oleosin 5' regulatory region operably linked to a fatty acid synthesis or lipid metabolism gene and regenerating a plant with increased levels of the product of said fatty acid synthesis or lipid metabolism gene.
Another aspect of the present invention provides methods of reducing levels of a product of a gene which is native to a plant which comprises transforming a plant cell with an expression vector comprising a subject oleosin regulatory region operably linked to a nucleic acid sequence which is complementary to the native plant gene. In this manner, levels of endogenous product of the native plant gene are reduced through the mechanism known as antisense regulation. Thus, for example, levels of a product of a fatty acid synthesis gene or lipid metabolism gene are reduced by transforming a plant with an expression vector comprising a subject oleosin 5' regulatory region operably linked to a nucleic acid sequence which is complementary to a nucleic acid sequence coding for a native fatty acid synthesis or lipid metabolism gene
The present invention also provides a method of cosuppressing a gene which is native to a plant which comprises transforming a plant cell with an expression vector comprising a subject oleosin 5' regulatory region operably linked to a nucleic acid sequence coding for the native plant gene. In this manner, levels of endogenous product of the native plant gene are reduced through the mechanism known as cosuppression. Thus, for example, levels of a product of a fatty acid synthesis gene or lipid metabolism gene are reduced by transforming a plant with an expression vector comprising a subject oleosin 5' regulatory region operably linked to a nucleic acid sequence coding for a native fatty acid synthesis or lipid metabolism gene native to the plant. Although the exact mechanism of cosuppression is not completely understood, one skilled in the art is familiar with published works reporting the experimental conditions and results associated with cosuppression (Napoli et al. 1990 The Plant Cell 2:270-289; Van der Krol 1990 The Plant Cell 2:291-299.
To provide regulated expression of the heterologous or native genes, plants are transformed with the chimeric gene constructions of the invention. Methods of gene transfer are well known in the art. The chimeric genes can be introduced into plants by leaf disk transformation-regeneration procedure as described by Horsch et al. 1985 Science 227:1229 . Other methods of transformation such as protoplast culture (Horsch et al. 1984 Science 223:496, DeBlock et al. 1984 EMBO J. 2:2143, Barton et al. 1983, Cell 32:1033) can also be used and are within the scope of this invention. In a preferred embodiment, plants are transformed with Agrobacterium-derived vectors such as those described in Klett et al. (1987) Annu. Rev. Plant Physiol. 38: 467. Other well-known methods are available to insert the chimeric genes of the present invention into plant cells. Such alternative methods include biolistic approaches (Klein et al. 1987 Nature 327:70), electroporation, chemically-induced DNA uptake, and use of viruses or pollen as vectors.
When necessary for the transformation method, the chimeric genes of the present invention can be inserted into a plant transformation vector, e.g. the binary vector described by Bevan, 1%. 1984 Nucleic Acids Res. 12:8711-8721. Plant transformation vectors can be derived by modifying the natural gene transfer system of Agrobacterium tumefaciens. The natural system comprises large Ti (tumor-inducing)-plasmids containing a large segment, known as T-DNA, which is transferred to transformed plants. Another segment of the Ti plasmid, the vir region, is responsible, for T-DNA transfer. The T-DNA region is bordered by terminal repeats, In the modified binary vectors, the tumor inducing genes have been deleted and the functions of the vir region are utilised to transfer foreign DNA bordered by the T-DNA border sequences. The T-regiön also contains a selectable marker for antibiotic resistance, and a multiple cloning site for inserting sequences for transfer. Such engineered strains are known as «disarmed» A. tumefaciens strains, and allow the efficient transfer of sequences bordered by the T-region into the nuclear genome of plants.
Surface-sterilized leaf disks and other susceptible tissues are inoculated with the «disarmed» foreign DNA-containing A. tumefaciens, cultured for a number of days, and then transferred to antibiotic-containing medium, Transformed shoots are then selected after rooting in medium containing the appropriate antibiotic, and transferred to soil. Transgenic plants are pollinated and seeds from these plants are collected and grown on antibiotic medium.
Expression of a heterologous or reporter gene in developing seeds, young seedlings and mature plants can be monitored by immunological, histochemical or activity assays. As discussed herein, the choice of an assay for expression of the chimeric gene depends upon the nature of the heterologous coding region, For example, Northern analysis can be used to assess transcription if appropriate nucleotide probes are available. If antibodies to the polypeptide encoded by the heterolog-ous gene are available, Western analysis and immunohistochemical localization can be used to assess the production and localization of the polypeptide. Depending upon the heterologous gene appropriate biochemical assays can be used. For example, acetyltransferases are detected by measuring acetylation of a standard substrate. The expression of a lipid desaturase gene can be assayed by analysis of fatty acid methyl esters (FAMES).
Another aspect of the present invention provides transgenic plants or progeny of these plants containing the chimeric genes of the invention. Both monocotyledonous and dicotyledonous plants are contemplated. Plant cells are transformed with the chimeric genes by any of the plant transformation methods described above. The transformed plant cell, usually in the form of a callus culture, leaf disk, explant or whole plant (via the vacuum infiltration method of Bechtold et al. 1993 C.R. Acad. Sci. Paris, 316:1194-1199) is regenerated into a complete transgenic plant by methods well-known to one of ordinary skill in the art (e.g. Horsch et al. 1985 Science 227:1129). In a preferred embodiment, the transgenic plant is sunflower, cotton, oil seed rape, maize, tobacco, Arabidopsis, peanut or soybean. Since progeny of transformed plants inherit the chimeric genes, seeds or cuttings from transformed plants are used to maintain the transgenic line.
The following, examples further illustrate the invention.

EXAMPLE 1: Isolation of Membrane-Bound Polysomal RNA and Construction of Borage cDNA Library

Membrane-bound polysomes were isolated from borage seeds 12 days post pollination (12 DPP)using the protocol established for peas by Larkins and Davies (1975 Plant Phys. 55: 749-756), > RNA was extracted from the polysomes as described by Mechler (1987 Methods in Enzymology 152: 241-248, Academic Press) Poly-A⁺ RNA was isolated from the membrane bound polysomal RNA using Oligotex-dT™ beads (Qiagen).
Corresponding cDNA was made using Stratagene's ZAP cDNA synthesis kit. The cDNA library was constructed in the lambda ZAP II vector (Stratagene) using the lambda ZAP II kit. The primary library was packaged with Gigapack II Gold packaging extract (Stratagene).

EXAMPLE 2 : Isolation of a Δ-6 Desaturase cDNA from Borage

Hybridization protocol

The amplified borage cDNA library was plated at low density (500 pfu on 150 mm petri dishes). Highly prevalent seed storage protein cDNAs were reduced (subtracted from the total cDNAs) by screening with the corresponding cDNAs.
Hybridization probes for screening the borage cDNA library were generated by using random primed DNA synthesis as described by Ausubel et al (1994 Current Protocols in Molecular Biology, Wiley Interscience, N.Y.) and corresponded to previously identified abundantly expressed seed storage protein cDNP.s. Unincorporated nucleotides were removed by use of a G-50 spin column (Boehringer Manheim). Probe was denatured for hybridization by boiling in a water bath for 5 minutes, then quickly cooled on ice. Nitrocellulose filters carrying fixed recombinant bacteriophage were prehybridized at 60°C for 2-4 hours in hybridization solution [4X SET (600 mM NaCl, 80 mM Tris -HCl, 4 mM Na₂EDTA; pH 7.8), 5X Denhardt's reagent (0.1% bovine serum albumin, 0.1% Ficoll, and (0.1% polyvinylpyrolidone), 100 µg/ml denatured salmon sperm DNA, 50 µg/ml polyadenine and 10 µg/ml polycytidine]. This was replaced with fresh hybridization solution to which denatured radioactive probe (2 ng/ml hybridization solution) was added. The filters were incubated at 60°c with agitation overnight. Filters were washed sequentially in 4X, 2X, and 1X SET (150 mM NaCl, 20 mM Tris-HCl, 1 mM Na₂EDTA; pH7.8) for 15 minutes each at 60°C. Filters were air dried and then exposed to X-ray film for 24 hours with intensifying screens at -80°C,
Non-hybridizing plaques were excised using Stratagene's excision protocol and reagents. Resulting bacterial colonies were used to inoculate liquid cultures and were either sequenced manually or by an ABI automated sequencer.

Random Sequencing of cDNAs from a Borage Seed 12 (DPP) Membrane-Bound Polysomal Library

Each cDNA corresponding to a non-hybridizing plaque was sequenced once and a sequence tag generated from 200-300 base pairs. All sequencing was performed by cycle sequencing (Epicentre). Over 300 expressed sequence tags (BSTs) were generated. Each sequence tag was compared to the GenBank database using the BLAST algorithm (Altschul et al. 1990 J. Mol. Biol. 215:403-410) . A number of lipid metabolism genes, including the Δ6-desaturase were identified.
Database searches with the cDNA clone designated mbp-65 using BLASTX with the GenBank database resulted in a significant match to the previously isolated Synechocystis Δ6-desaturase. It was determined however, that mbp-65 was not a full length cDNA. A full length cDNA was isolated using mbp-65 to screen the borage membrane-bound polysomal library. The resultant clone was designated pAN1 and the cDNA insert of pAN1 was sequenced by the cycle sequencing method. The amino acid sequence deduced from the open reading frame (Fig. 1, SEQ ID NO:1) was compared to other known desaturases using Geneworks (IntelligGenetics) protein alignment program. This alignment indicated that the cDNA insert of pAN1 was the borage □6-desaturase gene.

The resulting dendrogram (Figure 2) shows that Δ15-desaturases and Δ12-desaturases comprise two groups, The newly isolated borage sequence and the previously isolated Synechocystis Δ6-desaturase ( U.S. Patent No. 5,552,306 ) formed a third distinct group. A comparison of amino acid motifs common to desaturases and thought to be involved catalytically in metal binding illustrates the overall similarity of the protein encoded by the borage gene to desaturases in general and the Synechocystis Δ6-desaturase in particular (Table 1). At the same time, comparison of the motifs in Table 1 indicates definite differences between this protein and other plant desaturases. Furthermore, the borage sequence is also distinguished from known plant membrane associated fatty acid desaturases by the presence of a heme binding motif conserved in cytochrome b₅ proteins (Schmidt et al. 1994 Pant Mol. Biol. 26:631-642) (Figure 1). Thus, while these results clearly suggested that the isolated cDNA was a borage Δ6-desaturase gene, further confirmation was necessary. To confirm the identity of the borage □6-desaturase cDNA, the cDNA insert from pAN1 was cloned into an expression cassette for stable expression. The vector pBI121 (Jefferson et al. 1987 EMBO J. 6:3901-3907) was prepared for ligation by digestion with BamHI and EcoICR I (an isoschizomer of SacI which leaves blunt ends; available from Promega) which excises the GUS coding region leaving the 35S promoter and NOS terminator intact. The borage Δ6-desaturase cDNA was excised from the recombinant plasmid (pAN1) by digestion with BanxHI and xhoI. The XhoI end was made blunt by performing a fill-in reaction catalyzed by the Klenow fragment of DNA polymerase I. This fragment was then cloned into the BamHI/EcoICR I sites of pBI121.1, resulting in the plasmid pMN2.

TABLE 1

COMPARISON OF COMMOM AMINO ACID MOTIFS IN MEMBRANE-BOUND DESATURASES
Desaturase	Lipid Box	Metal Box	1	Metal Box 2
Borage	WIGHDAGH (SEQ. ID. NO:15)	HNAHH (SEQ. ID. NO:21)	FQIEHH (SEQ ID. NO:29)
Synechocystis	NVGNDANH (SEQ. ID NO:16)	HNYLHH (SEQ. ID. NO:22)	HQVTHX (SEQ. ID. NO:30)
Arab chloroplast	VLCHDCGH (SEQ, ID. NO:17)	HRTHE (SEQ. ID. NO:23)	HVIHH (SEQ. ID, NO:31)
Rice	VLGHDCGH (SEQ. ID. NO:17)	HRTHE (SEQ. ID. NO:23)	HVIHH. (SEQ. ID. NO:31}
Glycine chloroplase	VLGHDCGH (SEQ. ID. NO:17)	HRTHH (SEQ. ID- NO:23)	HVIHH (SEQ, ID. NO:31)
Arab. fad3	VLGHDCGH (SEQ. ID. NO:17)	HRTHH (SEQ. ID. NO:23)	HVIHH (SEQ. ID. NO:31)
Bressica fad 3	VLGHDCGH (SEQ. ID. NO:17)	HRTHH (SEQ. ID. NO:23)	HVIHH (SEQ. ID. NO:31)
Borage (P1-81)*	VIAHECGH (SEQ. ID. NO:18)	HRRHH (SEQ. ID NO:24)	HVAHH (SEQ. ID. NO:32)
Arab. fad2	VIAHZCGH (SEQ). ID. NO;18)	HRRHH (SEQ. ID. NO:24)	HVAHH (SEQ. ID. NO:321
Arab. chloroplast	VIGHDCAH (SEQ. ID. NO:19)	HDRHH SEQ. (ID. NO:25))	HIPHE (SEQ. ID. NO:33)
Glycine plastic	VIGHDCAH (SEQ. ID. NO:19)	HDRHH (SEQ. ID. NO:25)	HIPHH (SEQ. ID. NO:33)
Spinach plastidial n-6	VIGHDCAH (SEQ. ID. NO:19)	HDQHH (Seq. ID. NO:26)	HIPHH (SEQ. ID. NO:33)
synechocystisis	VVGHDCGH (SEQ. ID. NO:20)	HDHHH (SEQ. ID. NO:27)	HIFHH (SEQ. ID. NO:33)
Anabaena	VLGHDCGH (SEQ. ID. NO:17)	HNHEH (SEQ. ID. NO:28)	HVPHE (SEQ. ID. NO:34)

+P1-81 is & full length cDNA which was identified by EST analysis and shows high similarity to the Arbidopsis Δ12 desaturase (fad2)

EXAMPLE 3: Production of Transgenic Plants and Preparation and Analysis of Fatty Acid Methyl Esters (FAMEs)

The expression plasmid, pAN2 was used to transform tobacco (Nicotiana tabacum cv. xanthi) via Agrobacterium tumefaciens according to standard procedures (Horsch, et al. 1965 Science 227:1229-1231; Bogue et al. 1990 Mol. Gen. Genet. 221:49-57) except that the initial transformants were selected on 100 µg/ml kanamycin.
Tissue from transgenic plants was frozen in liquid nitrogen and lyophilized overnight. FAMEs were prepared as described by Dahmer, et al. (1989) J. Amer. Oil. Chem. Soc. 66: 543-548. In some cases, the solvent was evaporated again, and the FAMEs were resuspended in ethyl acetate and extracted once with deionized water to remove any water soluble contaminants, FAMEs were analyzed using a Tracor-560 gas liquid chromatograph as previously described (Reddy et al. 1996 Nature Biotech. 14:639-642).
As shown in Figure. 3, transgenic tobacco leaves containing the borage cDNA produced both GLA and octadecatetraenoic acid (OTA) (18:4 Δ6,9,12,15). These results thus demonstrate that the isolated cDNA. encodes a borage Δ6-desaturase.

EXAMPLE 4: Expression of Δ6-desaturase in Borage

The native expression of Δ6-desaturase was examined by Northern Analysis of RNA derived from borage tissues. RNA was isolated from developing borage embryos following the method of Chang at al. 1993 Plant Mol. Biol. Rep. 11:113-116. RNA was electrophoretically separated on formaldehyde-agarose gels, blotted to nylon membranes by capillary transfer, and immobilized by baking at 80°C for 30 minutes following- standard protocols (Brown T., 1996 in Current Protocols in Molecular Biology, eds. Auselbel, et al. [Greene Publishing and Wiley-Interscience, New York] pp. 4.9.1-4.9.14.). The filters were preincubated at 42°C in a solution containing 50%- deionized formamide, 5X Denhardt's reagent, 5X SSPE (900 mM NaCl, 50mM Sodium phosphate, pH7.7,- and 5 mM EDTA), 0.1% SDS, and 200 µg/ml denatured salmon sperm DNA. After two hours, the filters were added to a fresh solution of the same composition with the addition of denatured radioactive hybridization probe. In this instance, the probes used were borage legumin cDNA (Fig. 16A), borage oleosin cDNA (Fig. 16B), and borage Δ6-desaturase cDMA (pANI, Example 2) (Fig. 16C). The borage legumin and oleosin cDNAs were isolated by EST cloning and identified by comparison to the GenBank database using the BLAST algorithm as described in Example 2. Loading variation was corrected by normalizing to levels of borage EFla mRNA. EF1α mRNA was identified by correlating to the corresponding cDNA obtained by the EST analysis described in Example 2. The filters were hybridized at 42°C for 12-20 hours, then washed as described above (except that the temperature was 65°C), air dried, and exposed to X-ray film.
As depicted in Figs. 15A and 15B, Δ6-desaturase is expressed primarily in borage seed. Borage seeds reach maturation between 18-20 days post pollination (dpp). Δ6-desaturase mRNA expression occurs throughout the time points collected (8-20 dpp), but appears maximal from 10-16 days post pollination. This expression profile is similar to that seen for borage oleosin and 12S seed storage protein mRNAs (Figs. 16A, 16B, and 16C).

EXAMPLE 5: Isolation and Characterization of a Novel Oleosin cDNA

The oleosin cDNA (AtS21) was isolated by virtual subtraction screening of an Arabidopsis developing seed cDNA library using a random primed polymerase chain reaction (RP-PCR) cDNA probe derived from root tissue.

RNA PREPARATION

Arbidopsis thaliana Landsberg erecta plants were grown under continuous illumination in a vermiculite/soil mixture at ambient temperature (22°C). Siliques 2-5 days after flowering were dissected to separately collect developing seeds and silique coats. Inflorescences containing initial flower buds and fully opened flowers, leaves, and whole siliques one or three days after flowering were also collected. Roots were obtained from seedlings that had been grown in Gamborg B₅ liquid medium (GIBCO BRL) for two weeks. The seeds for root culture were previously sterilized with 50% bleach for five minutes and rinsed with water extensively. All tissues were frozen in liquid nitrogen and stored at until use. Total RNAs were isolated following a hot phenol/SDS extraction and LiCl precipitation protocol (Harris et al. 1978 Biochem. 17:3251-3256; Galau et al. 1981 J. Biol. Chem. 256:2551-2560). Poly A+ RNA was isolated using oligo dT column chromatography according to manufacturers' protocols (PHARMACIA or STRATAGENE) or using oligotex-dT latex particles (QIAGEN),

Construction of tissue-specific cDNA libraries

Flower, one day silique, three day silique, leaf, root, and developing seed cDNA libraries were each constructed from 5 µg poly A+ RN using the ZAP cDNA synthesis kit (Stratagene). cDNAs were directionally cloned into the EcoRI and XhoI sites of pBluescript SK(-) in the λ-EAPII vector (Short et al. 1988 Nucleic Acids Res. 16:7583-7600). Nonrecombinant phage plaques were identified by blue color development on NZY plates containing X-gal (5 bromo-4-, chloro-3-indoyl-β-D-galactopyranoside) and IPTG (isopropyl-1- thio-β- D-galactopyranoside). The nonrecombinant backgrounds for the flower, one day silique, three day silique, leaf, root, and developing seed cDNA libraries were 2.8%, 2%m 3.3%, 6.5%, 2.5%, and 1,9% respectively.

Random priming DNA labeling

The cDNA inserts of isolated clones (unhybridized cDNAs) were excised by EcoRI/XhoI double digestion and gel-purified for random priming labeling. Klenow reaction mixture contained 50 ng DNA templates, 10 mM Tris - HCl, pH 7. 5., 5 mM MgCl₂, 7,5 mM DTT, 50 uM each of dCTP, dGTP, and dTTP, 10 uM hexamer random primbers (Boehringer Mannheim), 50 µCi α-32 P-dATP, 3000 Ci/mmole, 10 mCi/ml (DuPont), and 5 units of DNA polymerase I Klenow fragment (New England Biolabs). The reactions were carried out at 37°C for one hour. Aliquots of diluted reaction mixtures were used for TCA precipitation and alkaline denaturing gel analysis. Hybridization probes were labeled only with Klenow DNA polymerase and the unincorporated dNTPs were removed using Sephadex R G-50 spin columns (Boehringer Mannheim).

Random Primed PCR

Double-stranded cDNA was synthesized from poly A+ RNA isolated from Arabidopsis root tissue using the cDNA Synthesis System (GIBCO BRL) with oligo dT12-18 as primers. cDNAs longer than 300 bp were enriched by Sephacryl S-400 column chromatography (Stratagene). Fractionated cDNAs were used as templates for RP-PCR labeling. The reaction contained 10 mM Tris-HCl, ph 9.0, 50 mM KCl, 0.1% Triton X-100, 2 mM MgCl2, 5 units Taq DNA polymeras (PROMEGA), 200 µM dCTP, cGTP, and dTTP, and different concentrations of hexamer random primers α-32P dATP 800 mci,/mmole, 10 mCi/ml (DuPont), and cold dATP in a final volume of 25 µl. After an initial 5 minutes at 95°C, different reactions were run through different programs to optimize RP-PCR cDNA conditions. Unless otherwise indicated, the following program was used for most RP-PCR cDNA probe labeling 95°C/5 minutes, then 40 cycles of 95°C 30 seconds, 18°C/1 second, ramp to 30°C at a rate of 0.1°C/second. 72°C/l minute. RP-PCR products were phenol/chloroform extracted and ethanol precipitated or purified by passing through Sephadex G-50 spin columns (Boehringer Mannheim).

Clone blot virtual subtraction

Mass excision of λ-ZAP cDNA libraries was carried out by co-infecting XL1-Blue MRF' host cells with recombinant phage from the libraries and ExAssist helper phage (STRATAGENE). Excised phagemids were rescued by SOLR cells. Plasmid DNAs were prepared by boiling mini-prep method (Holmes et al. 1981 Biochem. 114: 193-197) from randomly isolated clones. cDNA inserts were excised by EcoRI and XhoI double digestion, and resolved on 1% agarose gels. The DNAs were denatured in 0.5 N NaOH and 1.5 m NaCl for 45 minutes, neutralized in 0.5 M Tris-HCl, pH 8.0, and 1.5 M NaCl for 45 minutes, and then transferred by blotting to nylon membranes (Micron Separations, Inc.) in 10X SSC overnight, After one hour prehybridization at 65°C, root RP-cDNA probe was added to the same hybridization buffer containing 1% bovine albumin fraction V (Sigma), 1 mM EDTA, 0.5 M NaHP04, pH 7.2, 7% SDS. The hybridization continued for 24 hours at 65°C. The filters were washed in 0.5% bovine albumin, 1 mM: EDTA, 40 mM NaHPO4, pH 7.2, 5% SDS for ten minutes at room temperature, and 3 x 10 minutes in 1 mM EDTA, 40 mM NaHPO4, pH 7.2, 1% SDS at 65°C. Autoradiographs were exposed to X-ray films (Kodak) for two to five days at -80°C.
Hybridization of resulting blots with root RP-PCR probes «virtually subtracted» seed cDNAs shared with the root mRNA population. The remaining seed cDNAs representing putative seed-specific cDNAs, including those encoding oleosins, were sequenced by the cycle sequencing method, thereby identifying' AtS21 as an oleosin cDNA clone.

Sequence analysis of AtS21

The oleosin cDNA is 834 bp long including an 18 bp long poly A tail (Fig. 4, SEQ ID NO: 3) It has high homology to other oleosin genes from Arabidopsis as well as from other species, Recently, an identical oleosin gene has been reported (Zou, et al., 1996, Plant Mol. Biol. 31: 429- 433). The predicted protein is 191 amino acids long with a highly hydrophobic middle domain flanked by a hydrophilic domain on each side, The existence of two upstream in frame stop codons and the similarity to other oleosin genes indicate that this cDNA is full-length. Since there are two in frame stop codons just upstream of the first ATG, this cDNA is considered to be a full length cDNA (Figure 4, SEQ ID NO:3). The predicted protein has three distinctive domains based on the distribution of its amino acid residues. Both the N-terminal and C-terininal domains are rich in charged residues while the central domain is absolutely hydrophobic (Figure 5). As many as 20 leucine residues are located in the central domain and arranged as repeats with one leucine occurring every 7-10 residues. Other non-polar amino acid residues are also clustered in the central domain making this domain absolutely hydrophobic (Figure 6).
Extensive searches of different databases using both AtS21 cDMA and its predicted protein sequence identified oleosins from carrot, maize, cotton, rapeseed, Arabidopsis, and other plant species. The homology is mainly restricted to the central hydro-phobic domain. Seven Arabidopsis olesin sequences were found. AtS21 represents the same gene as Z54164 which has a few more bases in the 5' untranslated region. The seven Arabidopsis oleosin sequences available so far were aligned to each other (Figure 7). The result suggested that the seven sequences fall into three groups. The first group includes AtS21 (SEQ ID NO:5), X91918 (SEQ ID NO:6), and the partial sequence Z29859 (SEQ ID NO:7). Since X91918 (SEQ ID NO: 6) has only its last residue different from AtS21 (SEQ ID NO:5), and since Z29859 (SEQ ID NO:7) has only three amino acid residues which are different from AtS21 (SEQ ID NO:5), all three sequences likely represent the same gene. The two sequences of the second group, X62352 (SEQ ID NO:8) and Ato13 (SEQ ID NO:9), are different in both sequence and length, Thus, there is no doubt that they represent two independent genes, Like the first group, the two sequences of the third group, X91956 (SEQ ID NO:10) and L40954 (SEQ ID NO:11), also have only three divergent residues which may be due to sequence errors, Thus, X91956 (SEQ ID NO:10) and L40954 (SEQ ID NO:11) likely represent the same gene Unlike all the other oleosin sequences which were predicted from cDNA sequences, X62352 (SEQ ID NO:8) was deduced from a genomic sequence (Van Rooigen et al. 1992 Plant Mol Biol. 18:1177-1179). In conclusion, four different Arabidopsis oleosin genes have been identified so far, and they are conserved only in the middle of the hydrophobic domain.

Northern Analysis

In order to characterize the expression pattern of the native AtS21 gene, Northern analysis was performed as described in Example 4 except that the probe was the AtS21 cDNA (pAN1 insert) labeled with ³²P-dATP to a specific activity of 5 x 10⁸ cpm/ug.
Results indicated that the AtS21 gene is strongly expressed in developing seeds and weakly expressed in silique coats (Figure BA). A much larger transcript, which might represent unprocessed AtS21 pre-mRNA, was also detected in developing seed RNA. AtS21 was not detected in flower, leaf, root (Figure 8A.), or one day silique RNAs. A different Northern analysis revealed that AtS21 is also strongly expressed in imbibed germinating seeds (Figs. 13A and 13B)

EXAMPLE 6: Characterization of Oleosin Genomic Clones and Isolation of Oleosin Promoter

Genomic clones were isolated by screening an Arabidopsis genomic DNA library using the full length cDNA (AtS21) as a probe. Two genomic clones were mapped by restriction enzyme digestion followed by Southern hybridization using the 5' half of the cDNA cleaved by Sacl as a probe. A 2 kb SacI fragment was subcloned and sequenced (Fig. 9, SEQ ID NO:35). Two regions of the genomic clone are identical to the cDNA sequence. A 395 bp intron separates the two regions,
The copy number of AtS21 gene in the Arabidopsis genome was determined by genomic DNA Southern hybridization following digestion with the enzymes BamHI, EcoRI, HindIII, SacI and XbaI, using the full length cDNA as a probe (Figure BB). A single band was detected in all the lanes except SacI digestion where two bands were detected. Since the cDNA probe has an internal SacI site, these results indicated that AtS21 is a single copy gene in the Arabidopsis genome. Since it has been known that Arabidopsis genome contains different isoforms of oleosin genes, this Southern analysis also demonstrates that the different oleosin isoforms of Arabidopsis are divergent at the DNA sequence level.
Two regions, separated by a 395 bp intron, of the genomic DNA fragment are identical to AtS21 cDNA sequence. Database searches using the 5' promoter sequence upstream of AtS21 cDNA sequence did not identify any sequence with significant homology. Furthermore, the comparison of AtS21 promoter sequence with another Arabidopsis oleosin promoter isolated previously (Van Rooijen, et al., 1992) revealed little similarity. The AtS21 promoter sequence is rich in A/T bases, and contains as many as 44 direct repeats ranging from 10 bp to 14 bp with only one mismatch allowed. Two 14 bp direct repeats, and a putative ABA response element are underlined in Figure 9.

EXAMPLE 7: Construction of AtS21 Promoter/GUS Gene Expression Cassette and Expression Patterns in Transgenic Arabidopsis and Tobacco

Construction of AtS21 promoter/GUS gene expression cassette

The 1267 bp promoter fragment starting from the first G upstream of the ATO codon of the genomic DNA fragment was amplified using PCR and fused to the GUS reporter gene for analysis of its activity. The promoter fragment of the A1S21 genomic clone was amplified by PCR using the T7 primer GTAATACGACTCACTATAGGGC (SEQ ID NO:13) and the 21P primer GGGATCCTATACTAAAACTATAGAGTAAAGG (SEQ ID NO:14 ) complementary to the 5' untranslated region upstream of the first ATG codon (Figure 9). A BamHI cloning site was introduced by the 21P primer. The amplified fragment was cloned into the BamHI and SacI sites of pBluescript KS (stratagene) . Individual clones were sequenced to check possible PCR mutations as well as the orientation of their inserts. The correct clone was digested with BamHI and HindIII, and the excised promoter fragment (1.3 kb) was cloned into the corresponding sites of pBT101.1 (Jefferson, R.A. 1987a, Plant Mol. Biol. Rep. 5-387-405; Jefferson et al., 1987b, EMBO J. 6:3901-3907) upstream of the GUS gene. The resultant plasmid was designated pAN5 (Fig. 10). The AtS21 promoter/GUS construct (pAN5) was introduced into both tobacco (by the leaf disc method, Horsch et al., 1985; Bogue et al. 1990 Mol. Gen. Gen. 221:49-57) and Arabidopsis Colombia ecotype via vacuum infiltration as described by Bechtold, et al. (1993) C.R. Acad. Sci. Paris, 316: 1194-1199. Seeds were sterilized and selected on media containing 50 µg/ml kanamycin, 500 µg/ml carbenicillin. GUS activity assay: Expression patterns of the reporter GUS gene were revealed by histochemical staining (Jefferson, et al., 1987a, Plant Mol Biol. Rep. 5:387-405). Different tissues were stained in substrate solution containing 2 mg/ml 5-bromo-4-chioro-3-indolyl-β-glucuronic acid (X-Gluc) (Research Organics, Inc.), 0.5 mM potassium ferrocyanide, and 0.5 mM potassium ferricyanide in 50 mM sodium phosphate buffer, pH 7.0 at 37°C overnight, and then dehydrated successively in 20%, 40% and 80% ethanol (Jefferson, et al., 1987). Photographs were taken using an Axiophot (Zeiss) compound microscope or Olympus SZH10 dissecting microscope. Slides were converted to digital images using a Spring/Scan 35LE slide scanner (Polaroid) and compiled using Adobe Photoshop™ 3.0.5 and Canvas™ 3.5.
GUS activities were quantitatively measured by fluorometry using 2 mM 4-MUG (4-methylumbelliferyl-β-D-glucuronide) as substrate (Jefferson, et al., 1987). Developing Arabidopsis seeds were staged according to their colors, and other plant tissues were collected and kept at -80°C until use. Plant tissues were ground in extraction buffer containing 50 mM sodium phosphate, pH 7.0, 10 mM EDTA, 10 mM β-mercaptoethanol, 0.1% Triton X-100, and 0.1% sodium lauryl sarcosine. The tissue debris was removed by 5 minutes centrifugation in a microfuge. The supernatant was aliquoted and mixed with substrate and incubated at 37°C for 1 hour. Three replicas were assayed for each sample. The reactions were stopped by adding 4 volumes of 0.2 M sodium carbonate. Fluorescence was read using a TKO-100 DNA fluorometer (Hoefer Scientific Instruments). Protein concentrations of the extracts were determined by the Bradford method (Bio Rad).

Expression patterns of A1S21 promoter/GUS in transgenic Arabidopsis and tobacco

In Arabidopsis, GUS activity was detected in green seeds, and node regions where siliques, cauline leaves and branches join the inflorescence stem (Figures 11A and 11B). No GUS activity was detected in any leaf, root, flower, silique coat, or the internode regions of the inflorescence stem. Detailed studies of the GUS expression in developing seeds revealed that the AtS21 promoter was only active in green seeds in which the embryos had already developed beyond heart stage (Figures 11C and 11G). The youngest embryos showing GUS activity that could be detected by histochemical staining were at early torpedo stage. Interestingly, the staining was only restricted to the lower part of the embryo including hypocotyl and embryonic radical. No staining was detected in the young cotyledons (Figures 11D and 11E) . Cotyledons began to be stained when the embryos were at late torpedo or even early cotyledon stage (Figure 11F and 11H). Later, the entire embryos were stained, and the staining became more intense as the embryos matured (Figures 11I and 11J). It was also observed that GUS gene expression was restricted to the embryos. Seed coat and young endosperm were not stained (Figure 11C).
GUS activity was also detected in developing seedlings. Young seedlings of 3-5 days old were stained everywhere. Although some root hairs close to the hypocotyl were stained (Figure 11K), most of the newly formed structures such as root hairs, lateral root primordia and shoot apex were not stained (Figures 11L. and 11N). Later, the staining was restricted to cotyledons and hypocotyls when lateral roots grew from the elongating embryonic root. The staining on embryonic roots disappeared, No staining was observed on newly formed lateral roots, true leaves nor trichomes on true leaves (Figures 11M and 11N).
ALS21 promoter/GUS expression patterns in tobacco are basically the same as in Arabidopsis. GUS activity was only detected in late stage seeds and different node regions of mature plants. In germinating seeds, strong staining was detected throughout the entire embryos as soon as one hour after they were dissected from imbibed seeds. Mature endosperm, which Arabidopsis seeds do not have, but not seed coat was also stained (Figure 12A), The root tips of some young seedlings of one transgenic line were not stained (Figure 12B), Otherwise, GUS expression patterns in developing tobacco seedlings were the same as in Arabidopsis seedlings (Figures 12B, 12C, and 12D). Newly formed structures such as lateral roots and true leaves were not stained.

AtS21 mRNA levels in developing seedlings Since the observed strong activities of

AtS21 promoter/GUS in both Arabidopsis and tobacco seedlings are not consistent with the seed-specific expression of oleosin genes, Northern analysis was carried out to determine if AtS21 mRNA was present in developing seedlings where the GUS activity was so strong. RNAs prepared from seedlings at different stages from 24 hours to 12 days were analyzed by Northern hybridization using AtS21 cDNA as the probe. Surprisingly, AtS21 mRNl\ was detected at a high level comparable to that in developing seeds in 24-48 hour imbibed seeds. The mRNA level dropped dramatically when young seedlings first emerged at 74 hours (Figures 13A and 13B). In 96 hour and older seedlings, no signal was detected even with a longer exposure (Figure 13B). The loadings of RNA samples were checked by hybridizing the same blot with a tubulin gene probe (Figure 13C) which was isolated and identified by EST analysis as described in Example 2.
Since AtS21 mRNA was so abundant in seeds, residual AtS21 probes remained on the blot even after extensive stripping. These results indicated that AtS21 mRNA detected in imbibed seeds and very young seedlings are the carry-over of AtS21 mRNA from dry seeds. It has recently been reported that an oleosin Ato12 mRNA (identical to AtS21) is most abundant in dry seeds (Kirik, et al., 1996 Plant Mol. Biol. 31(2):413-417.) Similarly, the strong GUS activities in seedlings were most likely due to the carry-over of both β-glucuronidase protein and the de novo synthesis of β-glucuronidase from its mRNA carried over from the dry seed stage.

EXAMPLE 8 : Activity comparison between the AtS21 promoter and the 35S promoter

The GUS activities in transgenic Arabidopsis developing seeds expressed by the AtS21 promoter were compared with those expressed by the 35S promoter in the construct pBI221 (Jefferson et al. EMBO J. 6.3901-3907). The seeds were staged according to their colors (Table 2), The earliest stage was from globular to late heart stage when the seeds were still white but large enough to be dissected from the siliques. AtS21 promoter activity was detected at a level about three times lower than that of the 35S promoter at this stage. 35S promoter activity remained at the same low level throughout the entire embryo development. In contrast, AtS21 promoter activity increased quickly as the embryos passed torpedo stage and reached the highest level of 25.25 pmole 4-MD/min. µg protein at mature stage (Figure 5-8). The peak activity of the AtS21 promoter is as much as 210 times higher than its lowest activity at globular to heart stage, and is close to 100 times higher than the 35S promoter activity a.t the same stage (Table 2). The activity levels of the AtS21 promoter are similar to those of another Arabidopsis oleosin promoter expressed in Brassica napus (Plant et al. 1994, Plant mol. Biol. 25:193-205. AtS21 promoter activity was also detected at background level in leaf. The high standard deviation, higher than the average itself, indicated that the GUS activity was only detected in the leaves of some lines (Table 2). On the other hand, 35S promoter activity in leaf was more than 20 times higher than that in seed. The side by side comparisons of activities between AtS21 promoter and 35S promoter is shown in Figure 14.
Although the AtS21 promoter activity was about 3 times lower in dry seed of tobacco than in Arabidopsis dry seed, the absolute GUS activity was still higher than that expressed by the 35S promoter in Arabidopsis (Table 2). No detectable AtS21 promoter activity was observed in tobacco leaf (Figure 14).

Comparison of the AtS21 promoter versus the 35S promoter revealed that the latter is not a good promoter to express genes at high levels in developing seeds. Because of its consistent low activities throughout the entire embryo development period, 35S promoter is useful for consistent low level expression of target genes. On the other hand, the AtS21 promoter is a very strong promoter that can be used to express genes starting from heart stage embryos and accumulating until the dry seed stage. The 35S promoter, although not efficient, is better than the AtS21 promoter in expressing genes in embryos prior to heart stage.

TABLE 2

GUS ACTIVITIES OF AtS21 and 35S PROMOTER/GUS CONSTRUCTS
COLOR STAGE	WHITE G-H	WHITE/YELLOW H-T	YELLOW T-C	LIGHT GREEN EARLY C	DARK GREEN LATE C	GREEN/YELLOW/BROWN MATURE	BROWN DRY SEED	LEAF
AtS21	0.12±0.17	1.35±1.57	6.77±1.25	18.99±3.75	21.85±4.45	25.25±4.64	24.38±10.85	0.08±0.12
35S	0.30±0.06	0.25±0.08	0.29±0.04	0.28±0.03	0.33±0.06	0.26±0.04	0.31±0.02	6.56±0.74
AtS21	(In tobacco)						8.81±0.21	0.01±0.00

Abbreviations: G, globular stage; H, heart stage; T, torpedo stage; C, cotyledon stage. The GUS activities are in pmole 4-MU/µg protein.min. For AtS21 promoter the numbers are the average of five independent lines with standard deviations. Three repeats were assayed for each line. For 35S promoter the numbers are the average of three repeats of the same line with standard deviations.

EXAMPLE 9: Expression of the Borage Δ⁶-Desaturase Gene under the Control of the AtS21 Promoter and Comparison to Expression Under the Control of the CaMV 35S Promoter

In order to create an expression, construct with the AtS21 promoter driving expression of the borage Δ6-desaturase gene, the GUS coding fragment from pAN5 was removed by digestion with SmaI and EcoICR I. The cDNA insert of pAN1 (Example 2)was then excised by first digesting with XhoI (and filling in the residual overhang as above), and then digesting with SmaI, The resulting fragment was used to replace the excised portion of pAN5, yielding pAN3.

After transformation of tobacco and Arabidopsis following the methods of Example 7, levels of Δ⁶-desaturase activity were monitored by assaying the corresponding fatty acid methyl esters of its reaction products, γ-linolenic acid (GLA) and octadecatetraenoic acid (OTA) using the methods referred to in Example 3. The GLA and OTA. levels (Table 3) of the transgenic seeds ranged up to 6.7% of C18 fatty acids (Mean = 3.1%) and 2..8% (Mean = 1.1%), respectively. No GLA or OTA was detected in the leaves of these plants, In comparison, CaMV 35 S promoter/Δ⁶-desaturase transgenic plants produced GLA levels in seeds ranging up to 3,1% of CIS fatty acids (Mean = 1.3%) and no measurable OTA in seeds.

TABLE 3

EXPRESSION OF THE BORAGE Δ⁶-DESATURASE IN TRANSGENIC PLANTS
PROMOTER	PLANT	SEED				LEAF
		GLA*	RANGE	OTA*	RANGE	GLA	RANGE	GTA	RANGE
Cauliflower mosaic virus 3SS	tobacco	1.3	0.7-3.1	n.d		20	19-22	9.7	8-11
Arabidopsis oleosin	Arabidopsis	3.1	0-6.7	1.1	0-2.8	n-d.		n.d

*mean value expressed as the percent of the C₁₈ fatty acids n.d. not detected

EXAMPLE 10: Transformation of Oilseed Rape With an Expression Cassette Which Comprises the Oleosin 5' Regulatory Region Linked to the Borage Delta 6-Desaturase Gene

Oilseed rape, Cv. Westar, was transformed with the strain of Agrobacterium tumefaciens EHA105 containing the plasmid pAN3 (i.e. the borage Δ6 desaturase gene under the control of the Arabidopsis oleosin promoter-Example 9).
Terminal internodes of westar were co-cultivated for 2-3 days with induced Agrobacterium tumefaciens strain EHA105 (Alt-Moerbe et al. 1988 Mol. Gen. Genet. 213:1-8; James et al. 1993 Plant Cell Reports 12:559-563), then transferred onto regeneration medium (Boulter etal. 1990 Plant Science 70: 91- 99; Fry et al. 1987 Plant Cell Reports 6:321-325). The regenerated shoots were transferred to growth medium (Pelletier et al. 1983 1993 Mol.Gen. Menet. 191:244-250), and a polymerase chain, reaction (PCR) test was performed on leaf fragments to assess the presence of the gene.
DNA was isolated from the leaves according to the protocol. of KM Haymes et al. (1996) Plant Molecular biology Reporter 14 (3) :280-284, and resuspended in 100µl of water, without RNase treatment. 5µl of extract were used for the PCR reaction, in a final volume of 50µl, The reaction was performed in a Perkin-Elmer 9600 thermocycler, with the following cycles:

1 cycle: 95°C, 5 minutes
30 cycles: 95°C, 45 sec; 52°C, 45 sec
72°C, 1 minute
1 cycle: 72°C, 5 minutes

Fig. 1

5' TGG AAA TGG AAC CAT AA 3'
5' GGA AAC AAA TGA TOG TC 3'

Figure 17

The positive shoots were transferred to elongation medium, then to rooting medium (DeBlock et al 1989 Plant Physiol. 91: 694-701). Shoots with a well-developed root system were transferred to the greenhouse. When plants were well developed, leaves were collected for Southern analysis in order to assess gene copy number.
Genomic DNA was extracted according to the procedure of Bouchez et al. (1996) Plan Molecular Biology Reporter 14:115-123, digested with the restriction enzymes Bgl I and/or Cla I, electrophoretically separated on agarose gel (Maniatis et al. 1982, in Molecular Cloning; a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor/NY), and prepared for transfer to nylon membranes (Nytran membrane, Schleicher & Schuell) according to the instructions of the manufacturer. DNA was then transferred to membranes overnight by capillary action using 20XSSC (Maniatis et al. 1982). Following transfer, the membranes were crosslinked by UV (Stratagene) for 30 seconds and pre-hybridized for 1 hour at 65°C in 15 ml of a solution containing 6X.SSC, 0.5%SDS and 2.25% w/w dehydrated skim milk in glass vials in hybridization oven (Appligene), The membranes were hybridized overnight in the same solution containing a denatured hybridization probe radiolabelled with ³²P to a specific activity of 10⁸ cpm/µg by the random primer method (with the Ready-ToGo kit obtained from Pharmacia). The probe represents a PCR fragment of the borage delta 6-desaturase gene (obtained in the conditions and with the primers detailed above). After hybridization, the filters were washed at 65°C in 2XSSC, 0.1% SDS for 15 minutes, and 0.2XSSC, 0.1%SDS for 15 minutes. The membranes were then wrapped in Saran-Wrap and exposed to Kodak XAR film using an intensifying screen at -70°C in a light-proof cassette. Exposure time was generally 3 days.
The results obtained confirm the presence of the gene. According to the gene construct, the number of bands in each lane of DNA digested by Bgl I or Cla I represents the number of delta 6-desaturase genes present in the genomic DNA of the plant. The digestion with Bgl 1 and Cla 1 together generates a fragment of 3435 bp.

Claims

An isolated nucleic acid encoding an oleosin 5' regulatory region which directs seed-specific expression selected from the groups consisting of the nucleotide sequence set forth in SEQ ID NO: 12, the nucleotide sequence set forth in SEQ ID NO: 12 having an insertion, deletion, or substitution of one or more nucleotides, or a contiguous fragment of the nucleotide sequence set forth in SEQ ID NO:12.
An expression cassette which comprises the oleosin 5' regulatory region of Claim 1 operably linked to at least one of a nucleic acid encoding a heterologous gene or a nucleic acid encoding a sequence complementary to a native plant gene.
The expression cassette of Claim 2 wherein the heterologous gene is at least one of a fatty acid synthesis gene or a lipid metabolism gene.
The expression cassette of Claim 3 wherein the heterologous gene is selected from the group consisting of an acetyl-coA carboxylase gene, a ketoacyl synthase gene, a malonyl transacylase gene, a lipid desaturase gene, an acyl carrier protein (ACP) gene, a thioesterase gene, an acetyl transacylase gene, or an elongase gene.
The expression cassette of Claim 4 wherein the lipid desaturase gene is selected from the group consisting of a Δ6-desaturase gene, a Δ12-desaturase gene, and a Δ15-desaturase gene,
An expression vector which comprises the expression cassette of any one of Claims 2-5.
A cell comprising the expression cassette of any one of Claims 2-5.
A cell comprising the expression vector of Claim 6.
The cell of Claim 7 or 8 wherein said cell is a bacterial cell or a plant cell.
A transgenic plant comprising the expression cassette of any one of Claims 2-5.
A transgenic plant comprising the expression vector of Claim 6.
The plant of Claim 10 or 11 wherein said plant is at least one of a sunflower, soybean, maize, cotton, tobacco, peanut, oil seed rape or Arabidopsis plant.
Progeny of the plant of any one of claims 10 to 12.
Seed from the plant of any one of claims 10 to 12.
A method of producing a plant with increased levels of a product of a fatty acid synthesis gene or a lipid metabolism gene which comprises:
(a) transforming a plant cell with an expression vector comprising the isolated nucleic acid of Claim 1 operably linked to at least one of an isolated nucleic acid coding for a fatty acid synthesis gene or a lipid metabolism gene; and

(b) regenerating a plant with increased levels of the product of said fatty acid synthesis or said lipid metabolism gene from said plant cell.
The method of Claim 15 wherein said fatty acid synthesis gene or said lipid metabolism gene is at least one of a lipid desaturase, an acyl carrier protein (ACP) gene, a thioesterase gene an elongase gene, an acetyl transacylase gene, an acetyl-coA. carboxylase gene, a ketoacyl synthase gene, or a malonyl transacylase gene.
A method of decreasing production of a fatty acid synthesis or lipid metabolism gene in a plant which comprises:
(a) transforming a plant cell with an expression vector comprising the isolated nucleic acid of Claim 1 operably linked to a nucleic acid sequence complementary to a fatty acid synthesis or lipid metabolism gene; and

(b) regenerating a plant with decreased production of said fatty acid synthesis or said lipid metabolism gene.
A method of cosuppressing a native fatty acid synthesis or lipid metabolism gene in a plant which comprises:
(a) transforming a cell of the plant with an expression vector comprising the isolated nucleic acid of Claim 1 operably linked to a nucleic acid sequence encoding a fatty acid synthesis or lipid metabolism gene native to the plant; and

(b) regenerating a plant with decreased production of said fatty acid synthesis or said lipid metabolism gene.